1

ASYMMETRIC LOAD CARRYING WHILE WALKING ON A TREADMILL: GAIT KINEMATICS AND LOWER LIMB COORDINATION

By

JUNSIG WANG

A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2012

2

© 2012 Junsig Wang

3

To my family and friends, for all their support

4

ACKNOWLEDGMENTS

I would like to thank my advisor, Dr. Mark Tillman for the opportunity to study

Biomechanics under his guidance. Dr. Tillman’s knowledge, guidance, and wisdom

have allowed me to complete this project at this time. I would also like to thank my

committee members, Dr. Chris Hass and Dr. Christopher Janelle for your time and input.

Thank you to Ryan Roemmich and Jaimie Roper for all of their help and input as you

assisted me in Biomechanics lab throughout the past two years.

5

TABLE OF CONTENTS page

ACKNOWLEDGMENTS .................................................................................................. 4

LIST OF TABLES ............................................................................................................ 7

LIST OF FIGURES .......................................................................................................... 8

ABSTRACT ..................................................................................................................... 9

CHAPTER

1 INTRODUCTION .................................................................................................... 11

Specific Aims .......................................................................................................... 14 Hypotheses ............................................................................................................. 15

2 LITERATURE REVIEW .......................................................................................... 16

Human Locomotion and the Phases of Gait ........................................................... 16 Limb Coordination ................................................................................................... 19

Continuous Relative Phase ..................................................................................... 21 Influence of Asymmetrical Load-Carrying on Gait ................................................... 23

3 METHODS .............................................................................................................. 31

Participants ............................................................................................................. 31

Instrumentation and Task ....................................................................................... 31 Procedures ............................................................................................................. 32 Data Processing ..................................................................................................... 33

Gait Parameters: Hypothesis 1 ......................................................................... 34 Limb Coordination: Hypothesis 2 ...................................................................... 34

Statistical Analyses ................................................................................................. 35

4 RESULTS ............................................................................................................... 39

Gait Parameters: Hypothesis 1 ............................................................................... 39 Limb Coordination: Hypothesis 2 ............................................................................ 40

Intralimb Coordination ...................................................................................... 40 Interlimb Coordination ...................................................................................... 41

5 DISCUSSION ......................................................................................................... 50

Kinematic Gait Parameters: Hypothesis 1 .............................................................. 50 Limb Coordination: Hypothesis 2 ............................................................................ 55 Limitation of Study .................................................................................................. 60

6

LIST OF REFERENCES ............................................................................................... 67

BIOGRAPHICAL SKETCH ............................................................................................ 72

7

LIST OF TABLES

Table page 4-1 Comparisons of the loading conditions for gait kinematics ................................. 42

4-2 Comparisons of RMS difference for intralimb couplings ..................................... 43

4-3 Comparisons of RMS difference and CCC for interlimb couplings ..................... 44

8

LIST OF FIGURES

Figure page 2-1 Gait phases and subdivisions of stance and swing phase (Kirtley, 2008) .......... 29

2-2 Stride and step: A step is defined as the interval between heel contact of opposite feet. Two steps compose one gait stride .............................................. 29

2-3 Illustration of each segment angle and phase plot.............................................. 30

3-1 Illustration of four different load conditions ......................................................... 37

3-2 Bertec Instrumented Treadmill (Dual belt treadmill, Bertec, Columbs, Ohio, USA) ................................................................................................................... 38

3-3 Illustration of plug-in gait model in lower extremity with 16 lower extremity markers .............................................................................................................. 38

4-1 Effect of three different loading conditions (Figure 3-1) on stride length, cadence, and step width. *P< .05. ...................................................................... 45

4-2 Effect of three different loading conditions (Figure 3-1) on swing/stance ratio. *P< .05. ............................................................................................................... 46

4-3 RMS difference in thigh-shank coupling during both stance and swing phase for the loaded and unloaded sides. *P<. 05. ....................................................... 47

4-4 RMS difference in shank-foot coupling during both stance and swing phase for the loaded and unloaded sides. *P<. 05. ....................................................... 48

4-5 RMS difference and Cross-correlation coefficient for thigh-thigh coupling over a gait cycle. *P<. 05. ........................................................................................... 49

5-1 Mean CRP curves in thigh-shank on the loaded side during stance phase for baseline (no load) and each experimental condition (n=24). .............................. 62

5-2 Mean CRP curves in thigh-shank on the unloaded side during stance phase for baseline (no load) and each experimental condition (n=24). ......................... 63

5-3 Mean CRP curves in shank-foot on the loaded side during stance phase for baseline (no load) and each experimental condition (n=24). .............................. 64

5-4 Mean CRP curves in shank-foot on the loaded side during swing phase for baseline (no load) vs condition 1 and 3 (n=24). .................................................. 65

5-5 Mean CRP curves in thigh-thigh coupling over a gait cycle for baseline (no load) and each condition (n=24). ........................................................................ 66

9

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science

ASYMMETRIC LOAD CARRYING WHILE WALKING ON A TREADMILL:

GAIT KINEMATICS AND LOWER LIMB COORDINATION

By

Junsig Wang

August 2012

Chair: Mark Tillman Major: Applied Physiology and Kinesiology

Bags (backpacks and single sling/messenger bags) have become a daily

necessity for individuals of all ages. Carrying single strap bags induces asymmetrical

loading and could be harmful to the human body due to altered postures while walking.

The purpose of this study was to investigate the effect of different load carriage on gait

kinematics and coordinative patterns in the lower extremities. Twenty-four university

students participated in this study and walked on a treadmill at their preferred pace

under three different load conditions: condition 1 (5% of body weight in messenger bags

on each shoulder hanging vertically), condition 2 (10 % of body-weight in a messenger

bag on one shoulder hanging vertically) and condition 3 (10 % of body-weight in a

messenger bag on one shoulder with the bag draped across the trunk to opposite hip).

We examined gait kinematics (stride length, cadence, step width and swing/stance ratio)

and coordinative patterns in terms of intralimb and interlimb coupling utilizing

Continuous Relative Phase analyses (CRP). CRP was evaluated over three interlimb

(thigh-thigh, shank-shank, and foot –foot) and four intralimb couplings (thigh-shank and

shank-foot in both legs). The results were analyzed by using repeated measures one-

10

way Analyses of Variance (ANOVA) and indicated improved gait kinematics (step with

and swing/stance ratio) and coordinative patterns (thigh-shank, shank-foot and thigh-

thigh couplings) during bilateral carriage compared to unilateral load carriage.

Furthermore, asymmetric load carriage may have the potential to alter gait patterns and

coordinative mechanisms in the lower extremities which may be associated with the risk

of falling in taxing circumstances (e.g. slope, stairs, wet floors, etc.). Our results provide

a picture of the adaptive mechanisms utilized by the locomotor system under external

loading constraints.

11

CHAPTER 1 INTRODUCTION

People frequently carry loads using bags with shoulder straps. This allows

individuals to transport a variety of items and still have their hands free. Bags

(backpacks and single sling/messenger bags) have become a daily necessity for most

people and are utilized by individuals of all ages. Backpacks are used by soldiers,

hikers, and students. Students commonly use backpacks to carry their books,

computers, sports equipment, and other items on a daily basis (Negrini et al., 1999;

Grimmer et al., 1999; Grimmer & Williams, 2000; Whittfield et al., 2001). In recent years,

single strap bags have become more popular with students although they have been

used extensively by bicycle couriers for years and even by riders of the Pony Express

as early as 1860.

The majority of previous research related to carrying loads has been related to

conventionally worn backpacks. In fact, carrying symmetrically loaded backpacks using

both straps influences gait kinematics (Birrell & Haslam, 2009), posture and heartrate

(Simpson et al., 2011a), muscle activity patterns (Simpson et al., 2011b), balance (Pau

et al., 2011; May et al., 2009), and even decision making (May et al., 2009). Wearing

backpacks has also been suggested as a possible treatment for osteoporosis

(Wendlova, 2011) and camptocormia (Sakas et al., 2010). However, there is a relative

dearth of research concerning single strap bags that are associated with requisite

asymmetric loading.

Asymmetric loading occurs when load center of mass is displaced laterally in the

frontal plane (DeVita et al., 1991) which is observed when carrying a single sling or

messenger bag. This displacement of load center from body mid-line could affect the

12

gait pattern and cause back pain, as well as, injury of lower limbs and joints. Although

carrying backpacks over both shoulders could be the best way to prevent back pain

(Macias et al., 2008; Korovessis et al., 2005), students also choose asymmetrical load

carriage instead of symmetrical for a variety of reasons ranging from convenience to

fashion. In other words, carrying single strap bags are not conductive to the health of

individuals’ back, shoulder, or lower extremity and joints (Pascoe et al., 1997; Negrini et

al., 2007; Macias et al., 2008; Matsuo et al., 2008).

Students induce an asymmetric load by carrying their backpack with one strap on

one shoulder instead of both shoulders, concentrating the load on one side of the body.

This loading can place considerable stress and strain on the shoulder muscles (Marras

et al., 1997). It also seems that by carrying a single strap on one shoulder, individuals

must bend their trunk contralaterally to address the demand for dynamic balance.

Accordingly, increased trunk flexion toward the contralateral side (DeVita et al., 1991;

Pascoe et al., 1997) due to asymmetric load-carrying causes an increase in hip

abduction torque (DeVita et al., 1991; Matsuo et al., 2008). This indicates that

asymmetric loading alters dynamic balance when wearing a backpack with one strap

compared to both straps (Pascoe et al., 1997). In addition, the increased lateral bending

caused by asymmetric loading increases spinal stress (DeVita et al., 1991; Pascoe et

al., 1997; Legg & Cruz, 2004). The alterations of Contralateral trunk flexion during

walking may cause scoliosis or other permanent physical problems (Roaf, 1960; Hawes

& O’brien, 2006). As a result, increased bending stress on the trunk and spine may

exacerbate any abnormalities in the biomechanics of locomotion. These alterations

suggest that single strap bags may be a contributor to abnormal posture and motion of

13

the lower extremities in gait. Interestingly, the loads and mechanical adaptations

experienced when carrying loads tend to take place in the frontal plane. However, the

goal of load carriage during gait is to transport various items in a forward (sagittal plane)

direction. Therefore, most of the parameters examined in the current study were sagittal

plane variables. In addition, sagittal plane movements of the lower extremities are

typically the most meaningful measures compared to the other two planes in terms of

moving the body forward.

Physiological responses of single strap bags on gait measures can be seen

carrying golf bags. The muscle activation to maintain upright position when carrying the

single-strap of a golf bag on one shoulder, compared to double strap bags, may cause

higher demand on metabolic cost that induces fatigue (Ikeda et al., 2008). Asymmetric

load carriage is associated with higher oxygen uptake (Ikeda et al., 2008; Legg et al.,

1992) and heart rate (Ikeda et al., 2008) than backpack load carriage and could

increase muscle activation (Neumann et al., 1996) and fatigue. Fatigue caused by

single strap athletic bags could be one factor that may change posture and gait pattern.

In attempts to study posture and gait, previous researchers have modeled lower

extremity movement as a pendulum, with the segments of the leg oscillating during the

gait cycle. Coordinative structures have been useful tools to resolve and map each

segment pattern during gait. In movement science, examining how two oscillators are

coupled together during a limit cycle has been an important concern since human

movement involves coupling of multiple degrees of freedom in the human body.

Moreover, previous researchers have attempted to quantify coordinative patterns of two

oscillating segments during gait cycles by utilizing several methods in terms of

14

dynamical system theory (e.g. continuous relative phase, vector coding, and cross-

correlations). In this study, we used continuous relative phase to quantify and analyze

limb coordination of the lower extremities during the gait cycle. More specifically, we

adopted CRP to investigate effect of carrying messenger bag that affect inter limb and

intra limb coordination during the gait.

In sum, there has been a noticeable increase in single strap bags sales which

would indicate that more people are carrying them (Vicky, 2001). With the research on

numerous type of bags (schoolbags, military bags, athletic bags, and messenger bags)

showing that indeed, bags have a direct effect on the movement of the human body,

there seems to be a lack of research on other types of bags such as single strap bags.

In everyday life, bags can be considered necessities and therefore must be considered

if the movement is to be fully understood. A few researchers have investigated how

asymmetric load carrying affects the biomechanics of human movement. However, few

data exist regarding asymmetric load carriage and gait kinematics in terms of

coordination patterns of the lower extremities. Therefore, we attempted to show

evidence indicating that carrying single strap bags induces abnormal coordination

patterns in the lower body while walking.

Specific Aims

The aims of the research were 1) to investigate gait kinematics in response to three

different loading conditions and 2) to evaluate lower extremity limb coordination in

response to different loading conditions in university students during treadmill walking at

a self-selected pace. Baseline conditions were not included for gait kinematics because

the authors’ focus was to measure gait alterations in response to bilateral and unilateral

load carriage.

15

Hypotheses

The following general hypotheses and specific sub-hypotheses were tested.

1. Gait pattern would be altered when carrying two messenger bags (one on each shoulder) compared to a messenger bag on one shoulder.

A. Stride length would decrease and cadence would increase during symmetric loading compared to asymmetrical loading.

B. Step width would decrease during symmetric loading compared to asymmetric loading.

C. Swing/stance ratio would decrease on the loaded side and increase during symmetric loading compared to asymmetric loading.

2. Lower limb coordination (inter and intra limb coordination) would be altered when carrying two messenger bags (one on each shoulder) compared to a messenger bag on one shoulder.

A. RMS differences for intralimb (thigh-shank and shank-foot couplings) and interlimb (thigh-thigh, shank-shank and foot –foot couplings) would decrease during symmetric loading compared to asymmetric loading.

B. Cross-correlation coefficient values for intralimb and interlimb couplings would increase during symmetric loading compared to asymmetric loading.

16

CHAPTER 2 LITERATURE REVIEW

Single strap bags have become more popular with students in recent years, but

have been linked to altered gait kinematics (Pascoe et al., 1997). The proposed study

will be undertaken to examine the kinematic changes that may lead to variations in the

lower extremity movements and coordination. This chapter will be divided into four

sections. Section one will describe human locomotion and the phases of gait. Section

two will describe limb coordination during human gait. Section three will describe a

dynamical systems’ approach for analyzing the coordination of human movement.

Finally, section four will describe the influence that asymmetrical load-carrying has on

human locomotion as well as coordination patterns of lower extremity.

Human Locomotion and the Phases of Gait

Human gait is very complicated, involving continuous coordinated bipedal

locomotion. As a person moves forward, one of the lower limbs serves as a mobile

source of support while the other limb moves forward to become the new support site in

front of the current support site. To transfer body weight from one limb to the other, both

are alternately placed on the ground. This series of leg motions is called a “gait cycle”;

each gait cycle begins with ground contact of the heel and ends with the following heel

contact of the same limb (Figure 2-1). Several approaches have been proposed to fully

investigate the human gait cycle and numerous terminologies to define specific human

walking events have been introduced. In this chapter, we will look over the descriptive

phrasal terminology yielded from previous observational and kinematic analyses of

normal gait (Perry 1992; Sutherland et al., 1994).

17

Since this cycle of events is repeated by each limb with reciprocal motion timing

to the following part, there is no specific ending or starting point. However, for gait

analysis, an initial point of gait will be used to define the beginning of each gait cycle.

Each gait cycle is divided into two periods (stance and swing) based on heel contact

points of reciprocal foot motion (Figure 2-1). The swing phase is the time period in

which the foot is not in contact with the ground, starting with the point of toe off and

ending with the point of heel contact. The stance phase of the gait cycle is the time

period in which the foot is in contact with the ground. The stance phase begins with heel

contact, and ends with the point of toe-off.

The stance and swing phases are subdivided into three different intervals based

on the sequence of heel contact. The stance phase is subdivided into three smaller

phases: initial double support, single support, and terminal double support. The gait

cycle begins with initial double support phase when both feet are on the ground after

initial contact. When the opposite foot is lifted and starts to swing, single support

begins. In this phase, one leg is in contact with the ground to support the entire body

weight, while the other leg swings forward. When the swing leg begins a new heel

contact, the single support phase ends and terminal double support phase begins.

Terminal double support continues until the original stance limb is lifted to make a new

swing phase. The swing phase is also subdivided into three phases: initial swing, mid

swing, and terminal swing. The initial swing phase is the time from when the swing foot

leaves the ground to when it crosses the stance foot. In this phase, the foot is off the

ground and accelerating forward. Next, the mid-swing phase begins when the swing

limb is opposite the stance foot and ends when the tibia is vertical. In the mid-swing

18

phase, the swing foot moves forward and overtakes the limb in stance. Terminal swing

is from the time when the tibia is vertical until the foot makes contact with the ground.

Finally, the decelerating swing limb completes swing phase in preparation for the

stance phase.

In normal gait, the stance phase accounts for approximately 60% and the swing

phase occupies around 40% of the total gait cycle although the precise duration of the

intervals is altered according to the individual’s walking velocity (Mann, 1982). Also, the

time of the stance and swing phases of the gait cycle are inversely related to the

walking velocity in that both the stance and swing phases decrease in duration as

velocity increases (Mann, 1982). Likewise, when walking velocity slows, the times of the

stance and swing phases become longer.

Over the past several decades, numerous temporal technical terms describing gait

have been produced to help us scrutinize human gait. Moreover, a variety of gait

parameters has been codified to analyze and quantify gait. The terminology related to

human walking began with descriptive phrases obtained as a result of observational and

kinematic analysis of normal participants (Perry, 1992). For example, stride is the time

period from the heel strike of one foot to the following heel strike of the same foot. Steps,

regarded as the timing between two limbs, are indicated by the time period of initial

contacts of both feet. That is, a step is the time period from the heel strike of one foot to

the heel strike of the other foot. Two steps have been identified in one gait stride, since

the other initial contact begins at the midpoint of one stride (Perry, 1992, Figure 2-2).

The evaluation of walking is the primary tool for describing human gait patterns. In

the field of biomechanics, gait analysis is a method by which modern technologies are

19

used to incorporate information from a number of inputs to illustrate and analyze the

dynamics of gait (Perry, 1992). This basic approach to evaluating gait will be the focus

of the current research. It will provide evidence which will be used to address the

specific aim.

Limb Coordination

As previously stated, human gait is a complicated movement that includes cycles

and coordination of each segment. In bipedal locomotion, limb coordination is crucially

important and must be altered according to demands of varying external circumstances

(Reisman et al., 2005). To adapt to the demands of various environments, specific

coordinative patterns occur both within limb segments and between limbs. For example,

coordination of oscillating segments may change based on external restraints, such as

surface properties, or internal restraints, such as a neurological disease (Haddad et al.,

2005). Functional gait demands flexible limb movements to accommodate for different

external constraints (Reisman et al., 2005). Inter-segmental coordination demands

complicated interaction between the motor output of the neuromuscular system,

biomechanical factors, environmental factors, and task constraints (Higgins, 1985;

Kugler & Turvey, 1987). In response to the demands of a specific task, muscles must

be used in combination to bring about particular movements in several joints (Turvey,

1990). Therefore, movement coordination may involve both muscular action and

interactions at each joint. This approach to human movement may allow scientists to

implement important insight in their investigation of the neurological system. Thus,

human movement coordination has been defined as the neurobiological system’s ability

to generate complex movement incorporating several segments or joints (Forner-

Cordero et al., 2005). Coordinative inter-joint and inter-segmental movements are a

20

strategy of the central nervous system to control movement and maintain stability during

human locomotion (Kurz et al., 2004). Different coordination patterns have been

observed for various neurological disorders (i.e. hemiparesis, parkinson's disease,

stroke). In hemiparetic gait, an overactivated rectus femoris during swing phase was

associated with changing thigh-shank coordination (Forner-Cordero et al., 2005). Other

research suggests that asymmetric movement coordination in parkinson’s disease may

be involved in freezing of gait (Plotnik et al., 2005). Therefore, coordinative inter-

segmental movement results from individual muscles and neuropathways (Stergiou,

2004). Effective organization and modeling of gait coordination has been an important

focus for many researchers. Traditionally, biomechanical tools have been utilized to

analyze human movement patterns. However, scientists need novel approaches to gain

a more advanced understanding of the organization of human movement. The two

components of this approach are the multifactorial and nonlinear relationships of the

gait pattern (Stergiou, 2004). Moreover, mapping coordination patterns in limbs during

gait provides a clear mechanism of the neuromuscular system in low-dimensional terms

(Kurz et al., 2004). In a biomechanical system, to organize a coordinative model,

researchers have reduced the high available degrees of freedom in segmental

movement (Hamill et al., 1999). The concept of coordinative structure comes from

Bernstein’s work that indentified problems regarding the numerous degrees of freedom

and made efforts to simplify the many degrees of freedom in segmental movements and

reduce independent variables requiring control (Bernstein 1967; Turvey, 1990). In

conclusion, previous researchers studying gait coordination have been focused on

organizing the system for functional movement pattern, reducing numerous independent

21

variables and, summarizing the relationship between various components during gait.

Examining limb coordination patterns under external constraints (i.e. asymmetric load)

during gait may provide insight into specific control mechanisms and human movement

adaptations under various environmental conditions.

Continuous Relative Phase

Researchers have used different nonlinear dynamic techniques to investigate

variability in human movement based on the dynamical systems theory (Miller et al.,

2010). One prevalent dynamic systems analysis for studying coordination of segmental

movement can be evaluated through continuous relative phase (CRP). This measure

has been used to quantify the coordination between different body segments in several

activities (Varlet & Richardson, 2011). This approach to investigating the coordinative

pattern of gait is to recreate it as a dynamic system and study its stabilizing features.

Moreover, CRP indicates the aspects of interacting segments during the entire

movement cycle. Therefore, the relative phases of several oscillating segments can be

measured to quantify segmental coordination and evaluate movement patterns.

Prior to calculation for CRP, each segment angle (Figure 2-3) is normalized for each

trial using the following equation:

Also, angular velocity is normalized the following equation:

22

This is done by plotting the posing of a segment angle versus the angular velocity of

that segment in the “phase plot” (Hamill et al., 1999). The phase angles (Φ) are

obtained by calculating four-quadrant arctangent of the ratio of angular velocity by

angular position:

Then the continuous relative phase is calculated by subtracting the phase angle of the

proximal segment from that of the distal segment for a specific point during the gait

cycle.

Where continuous relative is the relative phase angle between the distal and proximal

segment, Φdistal segment (i.e., shank) is the phase angle of the distal segment, and

Φproximal segment (i.e., thigh) is the proximal segment. The main strength of the CRP

measure is that it compresses four variables (i.e., proximal and distal segments’ angles

and angular velocities). When the CRP is near 0°, the respective segments are in-phase,

while 180°of the CRP indicates that both segments are anti-phase (Haken et al.,1985;

Kelso et al., 1986; Scholz & Kelso, 1989; Diedrich & Warren, 1995). Positive relative

values indicate that the distal segment is ahead of the proximal segment, and negative

23

values indicate that the proximal segment is ahead in phase space (Clark & Phillips,

1993; Barela et al., 2000). The slope of the relative phase curve configuration indicates

which segment is moving faster during the period of the gait cycle (Barela et al., 2000).

A positive slope indicates that distal segment is moving faster in phase space, while a

negative slope indicates that the proximal segment is move faster in phase space.

Comparing the phase angles of two segments using relative phase provides insight of

normal and pathological gait patterns (Kurz et al., 2004).

Influence of Asymmetrical Load-Carrying on Gait

Because bag conditions are the main external constraints for students during

gait, the choice of bags can be a crucial factor in alterations of gait patterns. It seems

that by carrying single strap bags, abnormal gait patterns may be produced that have

the potential for inducing injuries in the lower extremity. Much of the research available

in the biomechanics field on asymmetrical gait has shown a detrimental influence on the

human body. Specifically, physical stress associated with carrying bags on one shoulder

altered the posture and gait pattern of children (Pascoe et al., 1997). Pascoe and

colleagues studied sixty-one youths, without bag, carrying a one-strap book bag, a two-

strap bag, and a one-strap athletic bag. In this study, the impact of carrying book bags

on static posture and gait kinematics was examined. This study utilized a two-phase

approach. In phase one, sixty-one students who transport their school materials had a

survey that provided descriptive and anthropometric characteristics, book bag weight,

typical bag carriage, and associated physical symptoms. In phase two, ten participants

were chosen for kinematic analysis based on the survey and underwent dynamic

walking at a self-determined pace. A 7.7kg (17% of body weight) book bag weight was

used in this study. A video analysis system was employed for digitizing the movements

24

for kinematic analysis. The results of the study indicated that the load of the book bags

resulted in a decrease in stride length and an increase in stride frequency and greater

angular motion of the head and the trunk was observed with the athletic bag trial.

Moreover, they showed carrying bags on one shoulder caused different movement

strategies such as lateral spinal bending and shoulder elevation while walking. They

also suggested that carrying backpacks promoted a significant forward lean of the head

and trunk compared to normal walking or one-strap athletic bags. In sum, they showed

clear evidence that carrying backpacks or messenger bags altered the posture and gait

of youths. However they did not evaluate the gait patterns of lower extremities related to

these abnormal postures.

Another study focusing on the effects of carrying an asymmetric sidepack on the

frontal plane joint moments in both lower extremities and in the L5/S1 joint during

walking was performed by Devita and Colleagues (1991). Five healthy males were

recruited for this study and required to walk under three load conditions: no load, a load

equivalent to 10% of the subject's body weight, and a load equivalent to 20% of the

subject's body weight. The participants were instructed to carry the pack on the left

shoulder while the subject’s left hand held the front of the strap. In this study, the

experiment was designed to obtain three-dimensional ground reaction force data and

frontal and sagittal plane film records for deriving lower extremity joint moments in both

plane and frontal plane L5/S1 moments. The results of this study showed that

asymmetric loads produced unbalanced frontal plane moments, large changes in right

limb moments and smaller changes in left limb moments and caused L5/S1 moment

dominance to shift to the right side during left and right single support phase. The

25

authors conclude that since the asymmetric load caused unilateral and unbalanced use

of trunk muscles on the non-loaded body side for both stance phases, asymmetric load

carrying can be a greater risk factor for injury than symmetric carrying (Devita et al.,

1991). Although they also showed different movement patterns such as unbalanced hip

and knee joint moments that may alter limb coordination, they did not report

coordination patterns in the lower body during asymmetric load carrying.

A separate study provided clear evidence that carrying single strap mailbags

alter the kinematics of the spine and induce stature loss (Fowler et al., 2006). They

used a specific mail bag (17.5% body weight load) to simulate the task of the postal

worker and attempt to quantify the effect of asymmetric load carriage on spinal

kinematics and stature loss. Thirteen markers were attached on the participant’s back

and one marker over each shoe. The changes in stature measure were assessed with a

stadiometer. The results of this study indicated increased forward leaning and bending

of the spine while carrying the messenger bag during walking. More specifically, they

showed up to five degrees of forward flexion in the thoracic spine in the sagittal plane

and up to a 12 increase in the lumbar region in the frontal plane. Stature loss was

evident in the loaded condition, producing a spinal shrinkage greater than the unloaded

condition (12.1±1.2mm loaded, 5.75 ±1.1mm unloaded). These findings provided strong

evidence that carrying asymmetric loads induces abnormal postures related to spinal

configuration during gait.

In addition, Hadded and colleagues (2005) performed a study on the coordinative

pattern in lower extremity to asymmetrical loading during walking. In this study twelve

healthy individuals were recruited to walk on a treadmill at their preferred walking speed.

26

Four different unilateral loads (0.9, 1.8, 2.7, 3.6, 4.5kg) for each lower leg were worn

2.5cm above the lateral malleolus utilizing a custom-made leg loading device. The

researchers hypothesized that this adaptation would alter interlimb coordination while

intralimb coordination would remain invariant. Continuous relative phase (CRP) was

employed to evaluate changes in limb coordination patterns. Changes in coordinative

patterns were quantified using both cross-correlation and root-mean-square (RMS)

difference measures. The result of this study indicated that increases in leg loads

resulted in RMS changes in the interlimb level of the loaded leg. At the interlimb level,

significant differences were found in both cross correlation and RMS measures (Haddad

et al., 2004). The authors of this study indicated that various adaptations in lower limb

coordination appear at both the intralimb and interlimb levels in response to unilateral

load changes. They suggested changes at the intralimb coordination level were greater

than the changes in interlimb coordination, in both RMS and cross-correlation measures.

More specifically, at the intralimb level (the thigh-shank and shank-foot couplings), the

RMS difference in the loaded leg increased significantly (indicating out of phase

movement) during both stance and swing phases. At the interlimb level, the RMS

differences significantly increased in all three interlimb couplings (thigh-thigh, shank-

shank, and foot-foot). Also, the cross-correlation coefficient systematically decreased

with greater loads. However they used a custom-made leg loading device that was

placed 2.5cm above the lateral malleolus. This load condition at the lower leg may

cause limited leg movement due to load itself and is difficult to apply to real life. In the

current study, single strap bags carried on one shoulder were used for enhancing

potential asymmetries instead of this leg load.

27

Yet another study investigated balance during asymmetric load-carrying and how

asymmetric loading affects lower limb coordination while walking. Five young women

and six elderly women walked under three different conditions: without a bag, with a 3kg

hand-held bag, and with an 8kg hand-held bag. The researchers also used CRP

measures to analyze lower limb coordination and CRP means for each load condition

were used to calculate the cross-correlation coefficients. The results of this study

showed that the mean of cross correlation coefficients for both interlimb and intralimb

level were close to 1.0 and each participant showed the same coordination pattern. The

author concluded that limb coordination is maintained under external constraints such

as asymmetric loads, thus hand-held load carrying minimally affected the coordinative

patterns in lower limbs (Matsuo et al., 2008). This study used 3kg and 8kg women’s

handbags as its load conditions but we used messenger bags to evaluate lower leg

movement in response to the diverse load conditions that are usually encountered by

university students.

Several studies have shown that asymmetrical load-carrying creates issues with

kinematics and coordinative patterns in lower extremity. Even though previous research

examined gait kinematic changes and coordinative patterns in lower extremity under

various loads conditions, little research has been done to show exactly how carrying a

single strap bag, compared to carrying a backpack, affects gait kinematics and lower

limb coordination for university students. Therefore it is important that more research be

conducted to investigate the effect of single strap bag carrying on coordination patterns

in lower extremity, providing insight into circumstances with a potential for injury. The

present study will be an attempt to simulate real life loading conditions utilizing different

28

types of bags. Furthermore, the evaluation of coordination patterns may reveal insight

regarding neuromuscular mechanisms related to gait.

29

Figure 2-1. Gait phases and subdivisions of stance and swing phase (Kirtley, 2008).

Figure 2-2. Stride and step: A step is defined as the interval between heel contact of opposite feet. Two steps compose one gait stride.

30

(A) (B)

Figure 2-3. Illustration of each segment angle and phase plot (A) Illustration of each segmental angle: thigh, shank, and foot angles in sagittal plane. (B) Phase plot illustrating phase angle based on angular displacement versus angular velocity over one gait cycle. Calculation of Phase angle(φ) of thigh and shank was obtained from arctangent function of angular velocity(ω)/ angular displacement (θ).

31

CHAPTER 3 METHODS

Participants

Twenty-four healthy students with an age range of 18 to 30 (12 males and 12

female; age 21.6±3.6 years; height 170.9±8.5; weight 67.2±12.5kg) participated in this

research. All participants were recruited from the University of Florida and surrounding

community by word of mouth. All were free of any pathology that would prevent them

from walking on a treadmill. Moreover, individuals were excluded if they had current

back or neck pain, any joint pain, history of arm pain, or spine pain in the previous three

months or if they were not healthy enough to handle the demands of the tests in this

study. Prior to participating in the study, each subject read and signed an informed

consent form approved by the university’s institutional review board. Twenty-two

participants were right-handed and during conditions 2 and 3 preferred to carry the

messenger bag on the right shoulder, while the other two left-handed individuals

preferred to carry the bag on left shoulder. Though both sides were loaded during

condition 1, the “loaded” and “unloaded” labels in the remaining portions of this

document refer to the sides which are loaded or unloaded during conditions 2 and 3,

respectively.

Instrumentation and Task

Two single strap bags (Figure 3-1) were utilized in this research to create three

different load conditions: Two single strap bags, one on each shoulder and hanging

down vertically (0% of body weight), two 5% body-weight single strap bags with one on

the right shoulder and one on the left (hanging vertically), and 10% body-weight single

strap bags in different positions while walking on a treadmill. The total weight carried in

32

the bags was normalized according to each subject’s body weight (10%BW) based on

existing literature as indicated in the following sentences. Previous studies using various

loads (10-20%) have shown the diverse effect on the human body during gait (Devita et

al., 1991; Fowler et al., 2005; Korovessis et al, 2005; Zultowski et al, 2008). In fact, the

bag load for college-age students has been recommended to not exceed 10-15% of the

individual’s body weight to prevent injuries and pains in the body (AN et al., 2010). Thus,

the load for the current study was determined based on these data and guidelines.

Furthermore, the messenger bags were positioned approximately 10 cm below the

ASIS makers across the conditions.

A Vicon motion analysis system with 7 high-resolution cameras (Vicon Nexus,

Oxford, UK) was used to collect three-dimensional kinematic data during each testing

condition. The cameras were positioned to record movement in the three cardinal

planes (frontal, sagittal, and transverse) during the gait cycle. Kinematic data were

captured at 120Hz. Calipers were used to record participants’ lower extremity

anthropometric measures and joint center locations were calculated from the static trial.

All participants walked on an instrumented treadmill (Figure 3-2) for each load condition.

The treadmill was built on two forceplates that allowed for measuring continuous ground

force data during gait. These data were used for identifying gait events (toe-off and

heel strike).

Procedures

After meeting the qualifications for participation, the body weights of participants

were measured to determine appropriate 10% body weight loads to be carried in the

different bag conditions. Participants were then prepared for retro-reflective marker

placement by changing into tight fitting shorts and shirts. Sixteen retro-reflective

33

markers were placed on the lower extremity over bony landmarks according to the

Vicon Plug-in-Gait marker system (Figure 3-3). All participants were instructed to walk

at a self-selected speed on the instrumented treadmill. To determine the "self-selected

speed" of each subject, the belt speed was initiated at 0.5 m/s. The speed was

gradually increased in increments of 0.1m/s until the participant signaled that his or her

preferred speed has been reached. Participants’ lower extremity movements were

captured digitally as they walked under different loading conditions. More specifically,

participants underwent testing in four different load conditions: no bag, carrying two

messenger bags on both shoulders with a 10%BW load, carrying the messenger bag on

one shoulder hanging vertically down to the hip with a 10%BW load, and carrying the

messenger bag on one shoulder draped across the trunk to opposite hip with a 10%BW

load (Figure 3-1). In condition 2 and 3, the messenger bags were worn on the side of

the preferred shoulder. They initially were asked to walk on a treadmill at their preferred

pace for five minutes in each condition, with the bag order randomly assigned (by using

a randomizing table). Thus, a total of four trials were completed (20 minutes of treadmill

walking). Each participant was allowed to rest as much as necessary between trials. We

collected data for the last one minute of each condition because of the large magnitude

of data necessary for processing. Thus, around 50 strides in each condition were

recorded.

Data Processing

As alluded to previously, three-dimensional lower extremity gait analysis was

performed on all participants. Kinematic data were processed and exported using the

Vicon nexus 1.7.1 software program (Vicon Nexus, Oxford, UK). Marker trajectories

were filtered using a fourth-order Butterworth filter, with a 10Hz cutoff frequency. Then,

34

kinetic data were used to identify the stance and swing phases of the gait cycle based

on each gait event (foot strike and toe off). Accordingly, the gait cycle was divided into

two phases: stance and swing.

Gait Parameters: Hypothesis 1

We examined gait kinematics (stride length, cadence, step width and swing/stance

ratio) while walking on a treadmill. Stride length was defined as the displacement of the

ankle marker along the walking axis during the time period from the heel strike to toe off

in the same foot (Reisman et al., 2005). Cadence was defined as the number of steps

during one minute. Step width was defined as the mediolateral distance from the swing

limb heel marker to the stance limb heel marker at each heel strike. Swing/stance ratio

was defined as the ratio between swing and stance time over a gait cycle. Around 50

gait cycles were evaluated for the gait kinematics using a custom Matlab code.

Limb Coordination: Hypothesis 2

Three segmental angles (thigh, shank, and foot), as well as, angular velocities

were assessed from raw kinematic data exported from the Vicon system (Figure 3-3).

Continuous relative phase (CRP) analyses were performed using Matlab. We focused

the kinematic analysis on CRP measures during the stance and swing phases of the

gait cycle, as well as, the gait cycle as a whole. Segmental angular velocities were

calculated from segmental angles in the sagittal plane utilizing the central difference

method. These data were then used to calculate phase angles from a phase plot, using

the arctangent of angular velocity / angular displacement at each data point.

For the intralimb couplings, CRP was calculated by subtracting the phase angle of the

proximal segment from that of the distal segment for each data point. More specifically,

the intralimb couplings were the thigh-shank and shank-foot in both limbs. For the

35

interlimb coupling, CRP was calculated by taking the difference between the phase

angles of both segments for each data point. The interlimb couplings were the thigh-

thigh, shank-shank and foot-foot. These procedures were described in more detail in

Chapter 2.

Thus, CRP was evaluated over three interlimb (thigh-thigh, shank-shank, and foot

–foot) and four intralimb couplings (thigh-shank and shank-foot in both legs) during the

gait cycle, as well as the swing and stance phases of the gait cycle. Since each trial had

a different length cycle, the data were processed using a linear length normalization that

allows the different trials to have equal data lengths for each gait cycle. In interlimb

coupling, the time series of one limb was time shifted to match left and right foot strikes.

Coordination patterns were quantified utilizing cross-correlation coefficient (CCC) and

root-mean-square (RMS) techniques. CCC was assessed by comparing the average

CRP in each load condition to the average CRP in the no-bag baseline condition during

gait cycle for interlimb and intralimb couplings. RMS difference was also evaluated by

comparing the average CRP in each load condition to the average CRP in the no bag

baseline condition. While the CCC measure indicates changes in the spatio-temporal

evolution of CRP patterns, RMS measures show information about the magnitude

differences in relative phase between the patterns (Haddad et al., 2004).

Statistical Analyses

Statistical analyses were performed using the SPSS® statistics (version 20; SPSS

Inc., Chicago, IL, USA). The effect of the different loading conditions on gait kinematics

(stride length, cadence, step width, and swing/stance ratio) and limb coordination

parameters were analyzed by using repeated measures one-way Analyses of Variance

(ANOVA). In addition, the influence of loading was further investigated for swing/stance

36

ratio and intralimb coupling by performing the statistical tests on the loaded and

unloaded limbs separately. Bonferroni’s posthoc procedure was used when appropriate.

The level of statistical significance for all tests was set at P<0.05.

37

Figure 3-1. Illustration of four different load conditions. Baseline (no load with one

messenger bag on each shoulder hanging vertically down to the hip), condition 1 (5% of body weight in messenger bags on each shoulder hanging vertically), condition 2 (10 % of body-weight messenger bag on one shoulder hanging vertically) and condition 3 (10 % of body-weight messenger bag on one shoulder with the bag draped across the trunk to opposite hip).

38

Figure 3-2. Bertec Instrumented Treadmill (Dual belt treadmill, Bertec, Columbs, Ohio, USA).

Figure 3-3. Illustration of plug-in gait model in lower extremity with 16 lower extremity markers.

39

CHAPTER 4 RESULTS

Gait Parameters: Hypothesis 1

Initially, we evaluated the effect of different loading conditions on gait kinematics

via stride length, cadence, step width and stance/swing ratio. A main effect for stride

length was detected (F (2, 46) = 5.418, P = .008). Follow-up testing revealed that stride

length during condition 2 was shorter than condition 3 (p =.006, Figure 4-1). No other

significant differences between conditions were observed (P >.05). Also, cadence varied

across conditions (F (2, 46) = 5.902, P = .005). More specifically, cadence during

condition 2 was significantly increased compared to condition 3 (P = .005; Figure 4-1).

No difference in cadence was observed in condition 1 versus 2 or condition 1 compared

to 3 (P >.05). A step width main effect was observed (F (2, 46) = 14.481, P < .001).

Post-hoc analysis revealed that the step widths for condition 1 compared to conditions 2

and 3 were significantly decreased (P = .002, P < .001; Figure 4-1). There was no other

significant difference between conditions for step width (P >.05). Finally, the

swing/stance ratio was assessed in both the loaded and unloaded sides over a gait

cycle. As mentioned previously, though both sides were loaded during condition 1, the

“loaded” and “unloaded” labels refer to the sides which were loaded or unloaded during

conditions 2 and 3, respectively. A significant main effect was observed in the loaded

limb (F (2, 46) = 14.274, P < .001). In particular, the swing/ stance ratio on the loaded

side during condition 1 was increased compared to conditions 2 and 3 (P < .001 and P

= .001, respectively; Figure 4-2). No significant difference was observed between

condition 2 and 3 (P > .05). Additionally, a main effect of swing/stance ratio on the

unloaded side was noted (F (2, 46) = 5.738, P = .006). Subsequent testing indicated

40

that the swing/stance ratios during condition 1 were significantly decreased compared to

conditions 2 and 3 (P = .014, P = .033; Figure 4-2). No difference was observed

between conditions 2 and 3 (P >.05). A summary of these comparisons appears in

Table 4-1.

Limb Coordination: Hypothesis 2

Intralimb Coordination

Intralimb couplings (thigh-shank and shank-foot) were examined for the loaded

and unloaded side during the stance and swing phase to investigate the mechanism of

coordinative patterns in the lower extremities over a given period. Coordinative patterns

were analyzed using RMS differences and Cross-Correlation Coefficients. RMS

differences in intralimb coupling were evaluated over stance and swing phases. A

significant main effect of RMS differences for thigh-shank coupling during the stance

phase in the loaded side was observed (F (2, 46) = 10.852, P < .001). Post-hoc testing

revealed that thigh-shank coupling during condition 1 was less than conditions 2 and 3

(P = .01, P = .001; Figure 4-3). Also, on the unloaded side, a main effect of RMS in

thigh-shank was detected (F (2, 46) = 5.378, P = .008). RMS changes during condition

1 were decreased compared to conditions 2 and 3 (P = .033 and P = .041, respectively;

Figure 4-3). However, no significant RMS difference was observed for thigh-shank

coupling during swing phase (P > .05).

For shank-foot coupling, significant differences in RMS were observed on the

loaded side during both stance (F (2, 46) = 9.220, P < .001) and swing phases (F (2, 46)

= 5.695, P = .006). RMS decreased during condition 1 compared to conditions 2 and 3

(P = .02 and P = .004, respectively, Figure 4-4). Also, during the swing phase, the RMS

difference during condition 1 in the loaded side was less than condition 3 (P = .027,

41

Figure 4-4). However, no significant differences of RMS in shank-foot coupling were

detected for the unloaded side (P> .05). Table 4-2 illustrates the loading conditions

compared to each other for RMS difference. Moreover, no cross-correlation coefficient

effects for thigh-shank and shank-foot coupling for either limb were displayed. All CCC

values for intralimb coupling were close to 1.

Interlimb Coordination

Interlimb coordination was examined via thigh-thigh, shank-shank and foot-foot

couplings. A significant main effect on RMS difference was observed in thigh-thigh

coupling (F (2, 46) = 5.276, P = .009). RMS differences during condition 1 in thigh-thigh

coupling were significantly less than condition 2 (P = .022; Figure 4-5). There were no

other significant differences between conditions (P >.05). No effect on RMS changes in

shank-shank or foot-foot was observed. Also Cross-Correlation Coefficient in thigh-thigh

varied across the conditions (F (2, 46) = 4.651, P = .014). CCC for thigh-thigh coupling

during condition 1 were significantly increased compared to condition 3 (P = .039;

Figure 4-5). There was no difference for CCC in condition 1 versus 2 or condition 1

versus 3 (P >.05). No effect on Cross Correlation Coefficient in shank-shank and foot-

foot coupling was displayed (P> .05). Table 4-3 illustrates the loading conditions

compared to each other for RMS difference and CCC in interlimb couplings.

42

Table 4-1. Comparisons of the loading conditions for gait kinematics. Kinematic variables Condition 1 v. 2 Condition 1 v. 3 Condition 2 v. 3

Stride Length

Cadence

Step Width

Swing/Stance Ratio-

Loaded

Swing/Stance Ratio-

Unloaded

Note. ↔ = no difference, ↑= increase, ↓= decrease

43

Table 4-2.Comparisons of RMS difference for intralimb couplings.

Note. ↔ = no difference, ↑= increase, ↓= decrease

RMS Difference Condition 1 v. 2 Condition 1 v. 3 Condition 2 v. 3

Loaded Thigh-shank

stance

swing

Shank-Foot

stance

swing

Unloaded Thigh-Shank

Stance

Swing

Shank-foot

Stance

Swing

44

Table 4-3. Comparisons of RMS difference and CCC for interlimb couplings. RMS difference Condition 1 v. 2 Condition 1 v. 3 Condition 2 v. 3

Thigh-Thigh

Shank-Shank

Foot-Foot

CCC

Thigh-Thigh

Shank-Shank

Foot-Foot

Note. ↔ = no difference, ↑= increase, ↓= decrease

45

Figure 4-1. Effect of three different loading conditions (Figure 3-1) on stride length,

cadence, and step width. *P< .05.

46

Figure 4-2. Effect of three different loading conditions (Figure 3-1) on swing/stance ratio.

*P< .05.

47

Figure 4-3. RMS difference in thigh-shank coupling during both stance and swing phase

for the loaded and unloaded sides. *P<. 05.

48

Figure 4-4. RMS difference in shank-foot coupling during both stance and swing phase

for the loaded and unloaded sides. *P<. 05.

49

Figure 4-5. RMS difference and Cross-correlation coefficient for thigh-thigh coupling

over a gait cycle. *P<. 05.

50

CHAPTER 5 DISCUSSION

Kinematic Gait Parameters: Hypothesis 1

The aim of this study was to investigate the effect of different loading conditions

on temporal-spatial gait patterns during treadmill walking for healthy university students.

Our first hypothesis was supported or partially supported for three of the five tests

performed related to gait kinematics (Table 4-1). More specifically, Gait pattern was

improved when carrying two messenger bags (one on each shoulder) compared to a

messenger bag on one shoulder (hanging vertically or draped across the body). Several

previous studies have identified the influence of loading on gait kinematics (Crosbie et

al., 1994; Pascoe et al., 1997; Cottalorda et al., 2003; Connolly et al., 2008; Birrel &

Haslam, 2009). Stride length and cadence in normal gait are approximately 1.41m and

113steps/min in adult participants, respectively (Perry, 1992). Our values regarding

stride length and cadence were less than those of normal walking; presumably because

participants in the current study walked on a treadmill. On a treadmill, individuals show

decreased stride length and faster cadence due to short treadmill belts (Cottalorda et al.,

2003). Accordingly, we observed lower cadence values across conditions during

treadmill walking than those of normal gait. In addition, the individuals tested here

selected a slower speed (0.8m/s) than normal walking velocity as their preferred walking

speed.

Crosbie and colleagues examined the effect of unilateral load carriage at 10% and

20% of body weight for twenty college students. They reported a decrease in step

length and an increase in cadence in response to unilateral carriage while walking

barefoot on a flat walkway (Crosbie et al., 1994). Pascoe et al. (1997) found that stride

51

length was decreased and cadence was increased when carrying a one strap bag, two-

strap bag, or a one-strap athletic bag compared to no bag. Additionally, these authors

did not find a significant difference between unilateral and bilateral load carriage. Our

results, in part, support this tendency regarding stride length and cadence and will be

discussed shortly. However, in the current study, the no load condition for gait

kinematics was not measured and thus was not available for comparison. In a more

recent study, Cottalorda et al. (2003) evaluated the effect of different methods of

backpack carrying (unilateral and bilateral carriage) while walking on a treadmill on gait

kinematics in children. Again, a decrease in stride length and an increase in cadence

when carrying a backpack on both shoulders compared to no bag was found, but no

difference between one strap and two straps was observed and only a single one strap

condition was evaluated. Furthermore, Connolly et al. (2008) studied thirty-two children

under two different load conditions while walking on an electronic walkway (GAITRite

system) and reported a decrease in stride length when carrying the backpack loaded

with 15% of body weight on one shoulder compared to those without a backpack.

However, they also reported no difference in stride length between two experimental

conditions (carrying a backpack on one shoulder and two shoulders). Our data were

consistent with these previous studies and did not show significant differences in stride

length and cadence between unilateral load carriage and bilateral load carriage.

Therefore, regardless of whether participants walked overground or on a treadmill,

stride length and cadence were not sensitive to symmetric and asymmetric loading. In

addition, it is possible that the mechanics of walking on a treadmill were externally

dictated. The velocity of the treadmill was fixed and may not have matched the subject’s

52

preferred walking speed in each condition. Thus, these parameters (stride length and

cadence) were not influenced by different methods of load carriage while walking on a

treadmill. However, we also evaluated the effect of different unilateral load carriage

strategies (conditions 2 and 3). When carrying the messenger bag on one shoulder

hanging vertically (condition 2), stride length was decreased compared to when the

participant carried a messenger bag on one shoulder with the bag draped across the

trunk to the contralateral hip (condition 3). This result was unexpected, could infer

potential benefits related to condition 3, and warrants further research. Changes in

stride length and cadence during walking are associated with the sagittal plane and may

contribute to gait stability. It has been shown that shorter stride length decreases

anterior-posterior stability (McAndrew & Dingwell, 2012). Moreover, faster cadence may

indicate an adaptive strategy to maintain dynamic balance by selecting a proper

cadence in response to external constraints (Arif et al., 2002; Rogers et al., 2008). None

of our participants had difficulty completing the walking tasks even though altered stride

length and cadence have been related to risk of falling (Maki, 1997). The shorter stride

length and faster cadence observed during condition 2 compared to condition 3 indicate

that this strategy used for unilateral load carriage could negatively alter walking

kinematics and increase risk.

In addition, we found altered gait parameters in the frontal plane. Step width during

asymmetrical load carriage (conditions 2 and 3) was increased compared to

symmetrical load carriage (condition 1). In general, step width in normal walking ranges

from 7.1 to 9.1cm (Murray et al., 1964). In the current study, all measurements of step

widths across the conditions were greater than those of normal gait. Several previous

53

researchers have examined different methods of carrying loads in terms of step width.

Crosbie et al. (1994) reported a significant decrease in step width for unilateral loading

conditions (10% and 20% of body weight) compared to a no load condition for male

college students and a non-significant trend towards an increase in step width for

female students. Connolly et al. (2008) observed decreased step widths while carrying a

10kg backpack on one shoulder compared to without a backpack for children. Also, a

more recent study showed a decrease in step width while carrying loads (10% and 20%

of body weight) in one hand compared to no load for healthy adults (Zhang et al., 2009).

As stated previously, in our current study, we did not evaluate the no load condition.

Thus, in this respect, our data could not be directly compared to previous research.

Despite this inconvenience, important inferences can be made. In general, unilateral

loads shift the center of mass towards the loaded side. Thus, increased step width

during asymmetrical loading conditions may indicate an increase in the base of support

boundaries to maintain lateral stability during unilateral carriage (Crosbie et al., 1994;

Connolly et al., 2008). Therefore, these alterations in step width may show a

compensatory adjustment of shifted COM in response to asymmetric load carriage. To

preserve dynamic balance, our participants may have needed the increased step width

to compensate for the laterally displaced COM.

We also found an effect of asymmetrical loading on stance and swing time over a

gait cycle. Normal gait consists of 62% stance phase and 38% swing phase, indicating

0.61 swing /stance ratio (Kirtley, 2008). In the current study, Swing/stance ratios were

measured in three different loading conditions for both loaded and unloaded sides and

were slightly below 0.61 for each. A higher swing/stance ratio indicates more swing time

54

and less stance time over a gait cycle and decreased stability (Berman et al., 1987;

Roth et al., 1997; Wu et al., 2000). On the loaded side, the swing/stance ratios for

asymmetrical loading (conditions 2 and 3) were increased compared to symmetrical

loading which indicates more stance time and less swing time over the gait cycle. For

the unloaded side, we found an opposite trend indicating that swing/stance ratios in

condition 2 and 3 were decreased compared to condition 1. Asymmetrical load carriage

affected both legs in a complementary manner because during one leg’s stance phase

the other leg has swing phase over a gait cycle. The effect of loads on swing/stance

ratio or stance and swing time has not been investigated in other studies. However, a

similar effect on swing and stance time has been seen for obese people (Błaszczyk et

al., 2011). These authors found 5% shorter swing and 3% longer stance for obese

individuals compared to lean participants. In the current study, the average effective

BMI for participants carrying a load of 10% of body weight was increased to 25.3. Thus

they were classified as overweight (BMI of 25-29.9) but not obese (BMI > 30). We found

similar changes on the loaded side which indicates more stance time to possibly avoid

losing dynamic stability. Moreover, complementary changes on the unloaded side may

be an additional adaptive strategy to maintain dynamic balance during gait. The

changes in both legs indicate an asymmetric gait pattern in terms of swing/stance ratio.

This change in symmetry relative to swing/stance ratio has been reported in hemiplegic

gait (Roth et al., 1997). Also, this temporal asymmetry was suggested in patients with

chronic stroke (Patterson et al., 2008). Thus, asymmetrical loading appears to cause

inverse swing/stance ratio changes in the loaded and unloaded sides. Although, these

alterations in symmetry assist to preserve dynamic balance, ultimately these changes

55

are less stable and this may lead to abnormal walking patterns in response to the

asymmetric load on one shoulder.

To conclude, gait patterns were altered by different loading conditions in the

current study (primarily in the frontal plane). Alterations in temporospatial parameters

during different load conditions may provide insight into adaptive strategies in response

to external constraints. These alterations may increase the risk for injury in some cases.

Most of the changes in gait patterns were detected in condition 2 compared to the other

conditions. Based on our findings, altered gait patterns observed while carrying a

messenger bag on one shoulder hanging vertically to the hip were needed to

compensate for decreased stability during gait.

Limb Coordination: Hypothesis 2

The present study of intralimb and interlimb coordination using CRP aimed to

evaluate coordinative lower extremity mechanisms in response to different loading

conditions during treadmill walking. Hypothesis 2 was supported by three out of four

dependent measures for stance and partially supported for two of four related to swing

in terms of RMS for intralimb coupling (Table 4-2). Also, hypothesis 2 was supported by

two out of six dependent measures related to interlimb couplings in terms of RMS and

CCC measures (Table 4-3). RMS difference and CCC were used to quantify differences

in intralimb and interlimb coordination for the three different loading conditions and

indicated that coordination was altered during unilateral load carriage. Specifically, two

average CRP curves in the baseline condition and each carriage condition were

compared to evaluate coordinative patterns for the three different loading conditions

using RMS and Cross-correlation coefficient. While the CCC measure indicates

changes in the spatio-temporal evolution of CRP patterns, RMS measures show

56

information about the magnitude differences in relative phase between the patterns

(Haddad et al., 2004).

For thigh-shank coupling, RMS difference in both loaded and unloaded sides

during stance phase were increased in condition 2 and condition 3 compared to

condition 1. We found no difference in RMS in the swing phase for thigh-shank coupling.

Although, we utilized different experimental conditions (bilateral and unilateral load

carriages) than those used by Haddad and colleagues (unilateral load carriage) we can

gain insight into the effect of unilateral load carriage. Specifically, similar tendencies for

RMS differences between no load and unilateral leg loads conditions have been

detected in the previous research (Haddad et al., 2004). They utilized six unilateral leg

loads (no load, 0.9, 1.8, 2.7, 3.6, and 4.5 kg) during treadmill walking and observed an

increase in RMS changes (4-9°) over both stance and swing phase as the leg load was

increased for the loaded limb in thigh-shank coupling with no effect on the unloaded

side. We also observed RMS changes on the unloaded side in stance and no difference

in RMS in swing phase in thigh-shank couplings. In part, their findings are in contrast

with our results due to different methods of carrying loads. However, RMS difference

only represents information on the magnitude of difference between two curves in

different loading conditions but does not show which curve is more out-of or in

phase. Complete CRP curves provide information regarding how in phase or out of

phase two segments are during the entire stance phase. In phase coupling indicates

that two segments of the body are temporally aligned regarding three components:

angular displacement, angular velocity and direction of two segments movement; non-

alignment is indicative of out-of phase coupling. This additional information provides

57

insight into the coordination patterns utilized in the different conditions tested here

(Figures 5-1 to 5-5). For example, the CRP pattern in condition 2 (Figure 5-1) was more

out-of phase than normal walking (baseline) in early stance phase (0-20%) for thigh-

shank coupling in the loaded side. Less out-of phase thigh-shank coupling was

observed on the loaded side for asymmetrical loads (conditions 2 and 3) compared to

the symmetrical load (condition 1).The majority of the RMS difference in condition 2 for

thigh-shank coupling on the loaded side was in mid to late stance (40-90%). Also, RMS

changes in condition 3 compared to condition 1 resulted from less out-of phase coupling

during stance (30-60% and 80-100%).These RMS changes indicate restricted thigh-

shank coupling in the loaded side during the unilateral load carriage on one shoulder.

We also found similar RMS modification in thigh-shank coupling on the unloaded side.

For the two asymmetric loading conditions, the CRP curves for condition 2 and

condition 3 (Figure 5-2) were less out-of phase for thigh-shank coupling compared to

condition 1. Major changes in condition 2 compared to condition 1 appeared in mid

stance (40-60%). Also, RMS changes in condition 3 compared to condition 1 appeared

in mid to late stance (30-100%). Less out-of phase thigh-shank patterns have been

seen in hemiparetic gait (Hutin et al., 2010). Moreover, Hutin and colleagues applied an

orthotic knee constraint in healthy participants, which produced a reduction in RMS

during swing and stance phases. The loss of thigh-shank coupling during stance phase

indicates one of the altered coordination patterns that may relate to restricted knee

flexion during gait. Therefore, asymmetrical loading may contribute to less out-of phase

movement, altering knee flexion during stance phase in both the loaded and unloaded

side compared to symmetrical loading.

58

For shank-foot coupling, we found significant RMS differences only on the loaded

side during both swing and stance phase. These RMS differences also represent partial

information. Further graphical analysis (Figure 5-3) revealed that the RMS differences in

condition 2 and 3 result from more out-of phase coupling compared to condition 1. The

CRP curves in conditions 2 and 3 during late stance phase (80-100%) show more out-of

phase movement compared to condition 1. These two asymmetrical loading conditions

contributed to intensified out of phase shank-foot coupling on the loaded side during late

stance (Figure 5-3). A significant RMS change in condition 3 on the loaded side

resulting from changes during early and mid-late swing phase (0-20% and 60-90%) was

also observed compared condition 1 which indicates a more out-of phase pattern in

condition 3 (Figure 5-4). Regarding intralimb coupling, we also detected CCC changes

in response to difference loading conditions. Our findings support previous research

(Haddad et al., 2004) that displayed invariance in CCC (all values were greater than

0.99) for the intralimb couplings while carrying a unilaterally applied leg load. Also,

these authors showed altered RMS for loaded limb coupling in response to increases in

leg load. However, we found an RMS difference in thigh-shank coupling on the

unloaded side as well as the loaded side. Also, Haddad et al. (2004) indicated only

RMS changes but not where these RMS changes resulted from. Another interesting

finding here is that CRP curves in thigh-shank and shank-foot display completely

different adaptations in response to asymmetric loading. Thigh-shank couplings on both

sides during stance were less out-of phase, while shank-foot coupling showed a more

out-of phase pattern on the loaded side during both swing and stance with an

asymmetrical load compared to symmetric loading. Thus, it is possible that increased

59

out-of phase shank-foot coupling may be related to increased ankle stiffness during

asymmetrical load carriage in the loaded limb which could result in restricted knee

movement.

The effect of asymmetric load carriage was also observed for interlimb

coordination. As mentioned before, RMS changes in thigh-thigh coupling while carrying

the asymmetric load in condition 2 were observed. More specifically, these changes

resulted from exaggerated asymmetry in thigh-thigh coupling during the gait cycle

(Figure 5-5). Also, the CCC value in thigh-thigh coupling was decreased when carrying

a messenger bag on one shoulder (condition 3), indicating a higher difference in

coordination compared to carrying two messenger bags (one on each shoulder).

Previous researchers reported an increase in RMS and a decrease in CCC regarding

interlimb couplings in response to unilateral leg loads during treadmill walking (Haddad

et al., 2004). They observed an increase in RMS and a decrease in CCC (.95-0.7) for

thigh-thigh, shank-shank and foot-foot couplings, as leg load was increased. However,

in the current study, increased asymmetry in interlimb coordination was found only for

thigh-thigh coupling in terms of RMS and CCC measures. In general, smoothness and

symmetry in gait are regarded as ‘normal walking’. Asymmetry during gait has been

observed in previous pathological gait research as stroke, Parkinson’s, hemiplegia and

cerebral palsy (Roth et al., 1997; Patterson et al., 2008; Johnsen et al., 2009; Meyns et

al., 2012). In these previous studies, researchers have focused on clinical treatments to

improve the asymmetry in gait for these patients. Asymmetry during gait, as observed

here, may be considered potentially injurious with symptoms beyond the capacity of

locomotor system.

60

The results of the current study suggest a variety of adaptations in intralimb and

interlimb coordination in response to symmetric and asymmetric load carriage. We

expected that altered coordinative patterns in the lower extremities would be displayed

during asymmetrical load carriage. As indicated before, changes in intralimb and

interlimb coordination were observed for the asymmetrical loading conditions compared

to symmetric loading condition. However, the two different asymmetrical conditions did

not show any difference in interlimb and intralimb couplings. We observed abnormal

patterns in thigh-shank, shank-foot, and thigh-thigh coordination. Based on our findings,

we recommend people avoid carrying a messenger bag on one shoulder (condition 2 or

3) in order to decrease abnormal limb coordination in daily activities. These alterations

may provide researchers with preliminary knowledge concerning diverse gait

adaptations caused by external constraints. The adaptations in limb coordination during

asymmetrical load carriage should be further investigated, as they may be indicative of

the possibility for acute and/or chronic joint injury and pain. Furthermore, asymmetric

load carriage may have potential to alter coordinative mechanisms in the lower

extremities which may be associated with the risk of falling in taxing circumstances (e.g.

slope, stairs, wet floors, etc.) Our results provide a picture of adaptive mechanisms of

locomotor systems under various constraints, which may enable us to educate at-risk

individuals (e.g., the elderly and children) on the dangers of abnormal gait patterns.

Limitation of Study

In our study, arm movement was constrained across the previously referenced

conditions because clear and accurate collection of kinematic data in the laboratory

environment required participants to cross their arms. However, arm movements are

indispensable parts of human locomotion that play an important role for contralateral

61

limb movements (Eke-Okoro et al., 1997). In addition, as alluded to previously, we

evaluated individuals walking at a fixed velocity on a treadmill. Thus, they could not alter

their walking velocity. However, changing walking speed in response to different loading

conditions could be an important adjustment to preserve dynamic balance during these

conditions. Subsequent work could involve calculating a preferred walking pace in each

experimental condition. Finally, the differences between overground and treadmill

walking should be considered. On a treadmill, altered gait patterns have been observed

because of short treadmill belts (Cottalorda et al., 2003). Thus, it is possible that

different coordinative patterns during overground walking may exist. Further, treadmill

walking may have limited application to real world conditions when people carry loads

for work or recreation.

62

Figure 5-1. Mean CRP curves in thigh-shank on the loaded side during stance phase for

baseline (no load) and each experimental condition (n=24).

Thigh-shank in loaded side

63

Figure 5-2. Mean CRP curves in thigh-shank on the unloaded side during stance phase for baseline (no load) and each experimental condition (n=24).

Thigh-shank in unloaded side

64

Figure 5-3. Mean CRP curves in shank-foot on the loaded side during stance phase for

baseline (no load) and each experimental condition (n=24).

Shank-foot in loaded side

65

Figure 5-4. Mean CRP curves in shank-foot on the loaded side during swing phase for baseline (no load) vs condition 1 and 3 (n=24).

Shank-foot in loaded side

sidn\ssssssswing

66

Figure 5-5. Mean CRP curves in thigh-thigh coupling over a gait cycle for baseline (no load) and each condition (n=24).

Thigh-thigh

67

LIST OF REFERENCES

An D-H, Yoon J-Y, Yoo W-G, Kim K-M. Comparisons of the gait parameters of young Korean women carrying a single-strap bag. Nurs Health Sci. 2010;12(1):87-93.

Arif M, Ohtaki Y, Ishihara T, Inooka H. Walking gait stability in young and elderly people and improvement of walking stability using optimal cadence. In: Micromechatronics and Human Science, 2002. MHS 2002. Proceedings of 2002 International Symposium on.; 2002:245 – 251.

Berman AT, Zarro VJ, Bosacco SJ, Israelite C. Quantitative gait analysis after unilateral or bilateral total knee replacement. J Bone Joint Surg Am. 1987;69(9):1340–1345.

Birrell SA, Haslam RA. The effect of military load carriage on 3-D lower limb kinematics and spatiotemporal parameters. Ergonomics. 2009;52(10):1298-1304.

Błaszczyk JW, Plewa M, Cieślinska-Swider J. Impact of excess body weight on walking at the preferred speed. Acta Neurobiol Exp (Wars). 2011;71(4):528–540.

Connolly BH, Cook B, Hunter S. Effects of backpack carriage on gait parameters in children. Pediatr Phys Ther. 2008;20(4):347–355.

Cottalorda J, Rahmani A, Diop M. Influence of school bag carrying on gait kinetics. J Pediatr Orthop B. 2003;12(6):357–364.

Crosbie J, Flynn W, Rutter L. Effect of side load carriage on the kinematics of gait. Gait & Posture. 1994;2(2):103–108.

DeVita P, Hong D, Hamill J. Effects of asymmetric load carrying on the biomechanics of walking. J Biomech. 1991;24(12):1119-1129.

Eke-Okoro ST, Gregoric M, Larsson LE. Alterations in gait resulting from deliberate changes of arm-swing amplitude and phase. Clin Biomech (Bristol, Avon). 1997;12(7-8):516–521.

Forner-Cordero A, Levin O, Li Y, Swinnen SP. Principal component analysis of complex multijoint coordinative movements. Biological Cybernetics. 2005;93:63-78.

Fowler NE, Rodacki ALF, Rodacki CD. Changes in stature and spine kinematics during a loaded walking task. Gait Posture. 2006;23(2):133-141.

Grimmer KA, Williams MT, Gill TK. The associations between adolescent head-on-neck posture, backpack weight, and anthropometric features. Spine. 1999;24(21):2262-2267.

Grimmer K, Williams M. Gender-age environmental associates of adolescent low back pain. Appl Ergon. 2000;31(4):343-360.

68

Haddad JM, van Emmerik REA, Whittlesey SN, Hamill J. Adaptations in interlimb and intralimb coordination to asymmetrical loading in human walking. Gait Posture. 2006;23(4):429-434.

Hamill J, van Emmerik RE, Heiderscheit BC, Li L. A dynamical systems approach to lower extremity running injuries. Clin Biomech (Bristol, Avon). 1999;14(5):297-308.

Hawes MC, O’brien JP. The transformation of spinal curvature into spinal deformity: pathological processes and implications for treatment. Scoliosis. 2006;1(1):3.

Hutin E, Pradon D, Barbier F. Lower limb coordination patterns in hemiparetic gait: factors of knee flexion impairment. Clin Biomech (Bristol, Avon). 2011;26(3):304–311.

Ikeda ER, Cooper L, Gulick P, Nguyen P. The metabolic cost of carrying a single- versus double-strap golf bag. J Strength Cond Res. 2008;22(3):974-977

Johnsen EL, Mogensen PH, Sunde NA, Ø stergaard K. Improved asymmetry of gait in Parkinson’s disease with DBS: gait and postural instability in Parkinson’s disease treated with bilateral deep brain stimulation in the subthalamic nucleus. Mov. Disord. 2009;24(4):590–597.

Kinoshita H. Effects of different loads and carrying systems on selected biomechanical parameters describing walking gait. Ergonomics. 1985;28:1347-1362.

Kirtley C. History of the study of locomotion: The modern era. From- http://www.univie.ac.at/cga/history/modern.html. 2008.

Korovessis P, Koureas G, Zacharatos S, Papazisis Z. Backpacks, back pain, sagittal spinal curves and trunk alignment in adolescents: a logistic and multinomial logistic analysis. Spine. 2005;30(2):247-255.

Kurz M and Stergiou N. Applied dynamic systems theory for the analysis of movement.

Legg SJ, Cruz CO. Effect of single and double strap backpacks on lung function. Ergonomics. 2004;47(3):318-323.

Legg SJ, Ramsey T, Knowles DJ. The metabolic cost of backpack and shoulder load carriage. Ergonomics. 1992;35(9):1063-1068.

Macias BR, Murthy G, Chambers H, Hargens AR. Asymmetric loads and pain associated with backpack carrying by children. J Pediatr Orthop. 2008;28(5):512-517.

Maki BE. Gait changes in older adults: predictors of falls or indicators of fear. J Am Geriatr Soc. 1997;45(3):313–320.

69

Mann R. Bioemchancis. In M. H. Jahss (Ed.), Disorders of the foot (pp. 37-67). Philadelphia: W. B. Saunders Company, 1982.

Marras W.S. and Granata K.P., Spinal loading during trunk lateral bending motion, J Biomech 30 (1997), pp. 697–703

Matsuo T, Hashimoto M, Koyanagi M, Hashizume K. Asymmetric load-carrying in young and elderly women: relationship with lower limb coordination. Gait Posture. 2008;28(3):517-520.

May B, Tomporowski PD, Ferrara M. Effects of backpack load on balance and decisional processes. Mil Med. 2009;174(12):1308-1312.

McAndrew Young PM, Dingwell JB. Voluntary changes in step width and step length during human walking affect dynamic margins of stability. Gait & posture. 2012.

Meyns P, Van Gestel L, Bruijn SM. Is interlimb coordination during walking preserved in children with cerebral palsy? Res Dev Disabil. 2012;33(5):1418–1428.

Miller RH, Chang R, Baird JL, Van Emmerik REA, Hamill J. Variability in kinematic coupling assessed by vector coding and continuous relative phase. J Biomech. 2010;43(13):2554-2560.

Murray MP, Drought AB, Kory RC. Walking Patterns of Normal Men. J Bone Joint Surg Am, 46, 335-360, 1964.

Negrini S, Carabalona R, Sibilla P. Backpack as a daily load for schoolchildren. The Lancet. 1999;354:1974.

Negrini S, Negrini A. Postural effects of symmetrical and asymmetrical loads on the spines of schoolchildren. Scoliosis. 2:8-8.

Neumann DA. Hip abductor muscle activity in persons with a hip prosthesis while carrying loads in one hand. Phys Ther. 1996;76(12):1320-1330.

Pascoe DD, Pascoe DE, Wang YT, Shim DM, Kim CK. Influence of carrying book bags on gait cycle and posture of youths. Ergonomics. 1997;40(6):631-641.

Patterson KK, Parafianowicz I, Danells CJ. Gait asymmetry in community-ambulating stroke survivors. Arch Phys Med Rehabil. 2008;89(2):304–310.

Pau M, Corona F, Leban B, Pau M. Effects of backpack carriage on foot-ground relationship in children during upright stance. Gait Posture. 2011;33(2):195-199.

Perry J. Gait analysis: Normal and pathological function. Thorofar, NJ: Slack Incorprated. 1992.

70

Plotnik M, Giladi N, Balash Y, Peretz C, Hausdorff JM. Is freezing of gait in Parkinson’s disease related to asymmetric motor function? Annals of Neurology. 2005;57(5):656-663.

Reisman DS, Block HJ, Bastian AJ. Interlimb Coordination During Locomotion: What Can be Adapted and Stored? Journal of Neurophysiology. 2005;94(4):2403 -2415.

Roaf R. Vertebral growth and its mechanical control. J Bone Joint Surg Br. 1960;42-B:40-59.

Rogers HL, Cromwell RL, Grady JL. Adaptive changes in gait of older and younger adults as responses to challenges to dynamic balance. J Aging Phys Act. 2008;16(1):85–96.

Roth EJ, Merbitz C, Mroczek K, Dugan SA, Suh WW. Hemiplegic gait. Relationships between walking speed and other temporal parameters. Am J Phys Med Rehabil. 1997;76(2):128–133.

Sakas DE, Boviatsis EJ, Stavrinou LC. Backpack treatment for camptocormia. Mov. Disord. 2010;25(13):2254.

Simpson KM, Munro BJ, Steele JR. Backpack load affects lower limb muscle activity patterns of female hikers during prolonged load carriage. J Electromyogr Kinesiol. 2011;21(5):782-788.

Simpson KM, Munro BJ, Steele JR. Effect of load mass on posture, heart rate and subjective responses of recreational female hikers to prolonged load carriage. Appl Ergon. 2011;42(3):403–410.

Stergiou N. Innovative Analyses of Human Movement. Human Kinetics; 2004.

Sutherland DH, Kaufman KR, and Moitoza JR. Kinematics of normal human walking. In: Rose J, and Gamble JG, Editors, Human walking, Williams & Wilkins, Baltimore, MD 1994: 23-45.

Varlet M, Richardson MJ. Computation of continuous relative phase and modulation of frequency of human movement. J Biomech. 2011;44(6):1200-1204.

Vicky Uhland. Bags carry the accoutrements of 21st Century life. Available at: http://www.wearablesmag.com/wearablesbusiness/mag/apparel_bags_carry_accoutrements/index.html.

Wendlova J. The importance of carrying a backpack in the rehabilitation of osteoporotic patients (biomechanical analysis). Bratisl Lek Listy. 2011;112(1):41-43.

Whittfield JK, Legg SJ, Hedderley DI. The weight and use of schoolbags in New Zealand secondary schools. Ergonomics. 2001;44(9):819-824.

71

Wu W-L, Su F-C, Cheng Y-M. Gait analysis after ankle arthrodesis. Gait & Posture. 2000;11(1):54–61.

Zhang XA, Ye M, Wang CT. Effect of unilateral load carriage on postures and gait symmetry in ground reaction force during walking. Comput Methods Biomech Biomed Engin. 2010;13(3):339–344.

Zultowski I, Aruin A. Carrying loads and postural sway in standing: the effect of load placement and magnitude. Work. 2008;30(4):359-368.

72

BIOGRAPHICAL SKETCH

Junsig Wang was born in Seoul, Korea. After receiving his Bachelor of Education

Degree in physical education from Kyung Hee University, he attended the University of

Florida for master’s degree in the fall of 2010. He received a Master of Science degree

in applied kinesiology and physiology (APK) with a concentration in Biobehavioral

Science from the University of Florida in August 2012.